🌊 👩‍🎨 😂 How to analyze competitors and set up a digital strategy in 40 minutes 🤶 ⭐️ ☝🏼

We spoke with Andrei Yunisov, Marketing Director at the Swiss company Bacula Systems. Previously, he worked with Beeline, MTS, Megafon, Tele2, Lamoda and other well-known brands. Andrey spoke about the SimilarWeb service and how competitor analysis will help build a strategy, plan a budget and make fewer mistakes.

This article is a synopsis of an interview with Andrei.

Tools that allow you to analyze competitors have appeared in Russia recently. I mean the numerical format. That is, not just go to the site and see how the pictures and texts are located, but to get serious numerical data.

Today we can find out what is happening on the site:

what traffic sources are used,
how are they used
what is their effectiveness and what intensity
what budgets are competitors spending (or not spending) on these channels.

The accuracy of this data has increased in the last 2 years, tools have appeared that help to solve these problems. Therefore, competitive analysis is the number one step in a digital project.

Website Analysis Tools

There are two types of devices that allow you to collect data. The first is the statistics counters on your site. But they can not be put on the site of competitors, if it is not hacked. The second type does not require any installation on the site - these are services such as SimilarWeb, ComScore and Alexa.

SimilarWeb uses several methods for retrieving data. The first is a partnership with toolbars. These are programs that are installed in the browser (for example, Yandex.bar). Once you have installed such a program in your browser, all your actions are recorded. You can see which sites you visit, in what order, how much time you spend there, what you do.

The second big source for analyzing programs is data from Internet providers. On their servers there is information what pages you visited, how much time you spent on them, etc.

For competitive analysis, use SimilarWeb

Today you can study competitors in 40 minutes: see what traffic sources they use, how users visit their site, where their bottlenecks are on the site. This is done using SimilarWeb.

SimilarWeb is a major service, and one of its nice bonuses is the free version. The functionality is enough to analyze any significant site. It is enough that he has more than 5 thousand visits per month. Anything less is not statistically representative.

For competitive analysis, use SimilarWeb.

Today you can study competitors in 40 minutes: see what traffic sources they use, how users visit their site, where their bottlenecks are on the site. This is done using SimilarWeb.

Previously, such tools were inaccurate due to lack of data. But in recent years, accuracy has improved, and in my experience, errors in the analysis are 20-30%.

Of course, we cannot take ideally accurate data, but for a competitive analysis it’s enough. We can understand where the maximum conversions are, where the maximum traffic comes from.

You can evaluate the trend: traffic is gaining on the competitor’s website or is falling. Accordingly, in dynamics we can draw conclusions about what strategy to implement.

What you can check in SimilarWeb

For example, you can check how many visitors were on a competitor’s site. You can watch up to 5 competitors at a time.

Here's what I check:

Look at the distribution of visitors from desktop and mobile devices. Find out how well a competitor has made a website for devices other than the desktop. Evaluate how your competitor uses mobile traffic for strategy. You can also see which competitors have a bounce rate for contextual advertising, and compare with your indicator.

That is, SimilarWeb reveals those competitor metrics that are in the standard statistics counters on your site.

Here is the main screen of the service. Here I chose three sites that we will analyze further:

The first is a relatively small online store artikstudio.ru. Compare it with Lamoda and Wildberries, two large clothing stores. Pay attention to how graphs and metrics behave when you analyze a small site and a large site. Or when you select a large site and compare with the same large competitor.

Above the main screen is the dynamics of attendance for 3 months. Below are engagement metrics: number of visits, average duration, depth, and failure rates.

I do not recommend making a conclusion based on averages - this is the "average temperature in the hospital." I advise you to look for places on the charts where the data behave abnormally or contradict each other. Analyze this particular place, and not just look at the averages at the bottom of the screen.

In the presentation above, I circled a red rectangle around the place where the contradiction was observed. Lamoda traffic decreased, while Wildberries rose.

When you look at your own data and see some sharp change, in most cases you understand what happened: the newsletter is gone, the budget for contextual advertising is increased, or the site crashes.

In the case of competitor analysis, you do not know this. But one must be able to put forward hypotheses. The marketer must be able to test these hypotheses.

If we have a situation like on the screen, then we can assume several things:

Lamoda could get hooked on the technical component;
Wildberries could launch an advertising campaign and win Lamoda customers;
the content has changed: a new range, prices or the appearance of a stock.

Everything must be checked and conclusions drawn. Understand if it works for you.

SimilarWeb — .

: , , — .

SimilarWeb

What should we do if we are inclined to believe that traffic has been affected by a change in content?

For example, we see that the anomaly fell in the last week of November, and now it's February. We can’t go to the competitor’s website and see the very content that brought traffic - everything has already changed there. Then we turn to the Wayback Machine service. This tool is also called a web archive:

Here we entered the Lamoda page and we see a timeline that shows pictures of the site at a particular point in time. If you click on the blue or green dots, then the appearance of the Lamoda site that was on this date will open.

Wayback Machine - a service that shows past versions of sites

Make it a rule: if traffic differs in SimilarWeb, you need to go to the Wayback Machine and see what happened there.

Wayback Machine is a bit like a search robot. He visits your site and saves a copy to himself. Sometimes this copy is not complete, so sites are not always displayed correctly.

This is a free project, it lives on user donations and on the team that programs it. Not all sites are saved there, but many.

You also need to understand that it stores quite a lot of data. Once with a friend we found what one site looked like in 1999. 20 years ago!

Learn as many metrics as possible in SimilarWeb

Look at the division of mobile traffic:

This is the data for 2016, only 30% of mobile traffic is here. In 2019, there will be more: about 40% of users will log in from the desktop and 60% from mobile devices.

If you see that the competitor’s site is accessed more from the desktop, then most likely the mobile version is your competitor’s weak point. Try investing in mobile traffic.

See bounce rates:

Blue and red lines are rising. This is an increase in the failure rate by about 10-15%. Here you can make assumptions. For example, if you see that traffic is growing along with this, then most likely competitors have included an advertising campaign with low-quality traffic.

See the distribution of traffic sources by resources:

Above are relative values (in percent), below are absolute values (in units of attendance.)

Here you can see what traffic sources your competitors use and in what volume. What is the distribution between them. You can see which traffic is more: organic or paid.

Here you can clearly see that Wildberries have much higher Brand Awareness - that is, users often go directly to this site, which is very good. So the brand gets free targeted visits to its resource.

Here you can click on the names of the sources under the graphs and get more details. For example, you can click on search traffic:

If you are looking for areas where you can relatively cheaply invade to conduct marketing in a crowded market, then look at the search distribution. In our case, it can be seen that two large stores have little traffic from video search. These are holes in the video search, which is not bad for us - you can pull traffic from there to your site quite inexpensively if you deal with your YouTube channel.

On social networks, there is also a numerical insight. It is clearly seen that most of the users come from VKontakte:

If we are engaged in SMM or targeted advertising, we need to somehow try to understand which platform we should invest in. VKontakte is believed to have more people, and so many decide to invest there. But it’s better to look at it from a different angle. For example, here you can see that Facebook has less pressure in the case of Lamoda and Wildberries. We can begin to do something targeted there. And take Instagram, naturally.

The next important report is popular pages:

Unfortunately, in the free version it is no longer available, only in the paid one.

But from it you can get a lot of interesting information. I took the Asos online store as an example. This report shows us the percentage of traffic for each page on the site.

At the 5th position, we see pgcheckout.aspx, which is related to the beginning of the order process. We see that 2.16% of users go through this page. What does this mean for us? This is a conversion marker. That is, we see that people who start to order something, only 2.16%.

On the 3rd place is the cart page. That is, we know that 3.11% falls into the basket, and 2.16% falls on the checkout. And if the address of each next step of the basket is different, then we can evaluate all the stages of the conversion of this store.

Immediately, we see that the sale section (at number 9) occupies half the total traffic of the women section (line 2). Then you can look at signin (line 4) and understand that 2.78% of people log in to the site.

In fact, you can do usability analysis on competitor data and evaluate the functionality.

The paid version of SimilarWeb costs around a few hundred dollars - not such a terrible amount.

And the last useful feature for today is the Outbound Links report. It shows which sites users go to after viewing. I checked Wildberries:

On the 4th place we see the site wbpay.ru. We look at it, and it becomes clear that this is the site through which Wildberries customers pay for purchases. We get:

So using SimilarWeb, you can study competitors in 40 minutes: traffic sources, user behavior, functionality, potential errors. If you look in more detail and combine SimilarWeb with other tools, it will be easier to estimate the budget for traffic sources.